Ž .
Ž .
amplified fragment length polymorphism DNA markers. The female Om parent had a total of Ž
. Ž
78 segregating markers 17 microsatellites, 61 AFLPs . Of these, 62 13 microsatellites, 49 .
Ž .
AFLPs were linked in 14 linkage groups covering a total of 514 centimorgans cM . The first Ž
. Ž
generation F hybrid male parent had a total of 229 segregating markers 62 microsatellites, 167
1
. Ž
. AFLPs , of which 214 60 microsatellites, 154 AFLPs were linked in 24 linkage groups covering
a total of 1632 cM. The construction of these maps is a key step in a molecular marker-assisted Ž
. breeding program to detect quantitative trait loci QTL for cold and salinity tolerance and carcass
quality in tilapia. q 2000 Elsevier Science B.V. All rights reserved.
Keywords: Tilapia; Oreochromis; Genetic map; AFLP; Microsatellite; Cichlidae; Interspecific hybridization
1. Introduction
Ž .
The use of interspecific composite or complex crosses, an established practice in plant breeding, is a means of achieving wide genetic and phenotypic variability. The
Ž .
method is based on the creation of an artificial center of origin ACO via composite interspecific crosses. The ACO contains wide genetic diversity and presents opportuni-
ties for genes to recombine and interact with other genes originating from different species; combinations and interactions that are impossible in any of the pure species. An
Ž .
adaptation of this method, termed multiple re-speciation MRS , recently was applied to Ž
. develop new cultivars of carnations Umiel, 1993 .
We have constructed a tilapia ACO, exploiting the relative ease of making interspe- Ž
. cific hybrids with these species Wohlforth and Hulata, 1983 . The ACO involved four
Ž Ž
. Ž
. .
tilapiine species: Oreochromis niloticus wild-type On and red ROn strains , O. Ž
. Ž
. Ž
. aureus Oa , O. mossambicus Om , and Sarotherodon galilaeus Sg .
Genetic maps have been constructed for a number of fish species including zebrafish, Ž
. Danio rerio
Postlethwait et al., 1994; Johnson et al., 1996 ; rainbow trout, On- Ž
. Ž
corhynchus mykiss Young et al., 1998 ; and Nile tilapia, O. niloticus Kocher et al., .
1998 . In order to best utilize the unique genetic material produced in the creation of the ACO, we are producing linkage maps from several three-way and four-way crosses
Ž .
3WC and 4WC, respectively produced as part of the breeding program. These maps provide linkage information for markers that are segregating directly in the multi-specific
Ž .
crosses, they further develop the existing map of the On genome Kocher et al., 1998 , Ž
. and they establish a framework for the detection of quantitative trait loci QTL for traits
of importance to the culture of tilapia. These maps will also be used to evaluate the genetic composition of the strains derived from the ACO. Here we report on the results
of the breeding program that created the ACO and linkage maps from a 3WC family containing genetic material of Oa, Om, and ROn.
2. Materials and methods
2.1. Breeding program Four different species of tilapia were used for producing the composite population
Ž .
Table 1 . Breeding of tilapias was carried out at the Department of Aquaculture,
Table 1 Tilapia species and strains used in the breeding program of this study
Species Abbreviation
Origin Ž
. Ž
. O. aureus Steindachner
Oa Local Israeli Mehadrin strain, described by Hulata et al. 1993
Ž .
Ž .
O. niloticus Linnaeus On
Ghana strain Mires, 1977; Hulata, 1988 Ž
. Red O. niloticus
ROn Originated in Lake Manzala, Egypt McAndrew et al., 1988 ,
and obtained from the Institute of Aquaculture, University of Stirling, Scotland
Ž .
Ž .
O. mossambicus Peters Om
Natal, South Africa Hulata, 1988 Ž
. Ž
S. galilaeus Linnaeus Sg
Endemic species originating in Lake Tiberias Kinnereth, Sea of . Ž
. Galilee
Ben-Tuvia, 1959; Goren, 1974
Agricultural Research Organization, Bet Dagan, Israel, except where otherwise stated. Parental fish were individually tagged using colored and numbered plastic discs,
attached to the fish on nylon threads pushed through the dorsal muscle and knotted. Ž
Parents of all crosses retained for further study were PIT tagged Destron-Fearing, St. .
Paul, MN, USA and placed in holding tanks. Starting with the parental species,
Ž .
Ž .
Ž .
two-way crosses 2WC were made to produce first F and second F
generation
1 2
populations. Some of the 2WCs were crossed to a third species to produce a series of 3WC, or intercrossed to create 4WC.
Reproduction procedures were similar to those described in Rosenstein and Hulata Ž
. 1994 , except that eggs removed from the buccal cavities of females were transferred to
hatching jars to complete incubation. The S. galilaeus = Oreochromis sp. F hybrids
1
Ž were produced by artificial fertilization Don and Avtalion, 1986; Yeheskel and Avtal-
. ion, 1988 at the Laboratory of Fish Immunology and Genetics, Bar-Ilan University,
Ramat Gan, Israel. To produce the artificially fertilized groups, a batch of eggs of a single S. galilaeus female was divided into four sub-samples, each fertilized with sperm
collected from males of the four Oreochromis stocks. Samples of the F progeny were
1
transferred to Bet Dagan for use in the breeding program. 3WCs and 4WCs were obtained from these F fish by natural spawning.
1
2.2. Linkage mapping 2.2.1. Tissue
Sixty-three 3WC progeny of a single mating between an Om female and an F
1
Ž .
Oa = ROn male were chosen for evaluating joint segregation of microsatellite and Ž
. Ž
2
. amplified fragment length polymorphism AFLP markers. Fin tissue 1–4 cm
was collected from adult fish of each parental species, including all parents of the progeny
group contributing to the ACO, and stored in 95 ethanol. The tissue of progeny was Ž
. obtained from newly hatched fry 10–14 days old and placed in 95 ethanol for
storage. DNA from a 0.7-cm
2
piece of fin or an entire fry was extracted using the Ž
. CTAB phenolrchloroform protocol of Saghia-Maroof et al. 1984 and Doyle and Doyle
Ž .
Ž .
1987 , as modified by Grewe et al. 1993 .
2.2.2. Microsatellites Ž
. Fluorescently labeled microsatellite primer pairs n s 133 developed from an On
Ž .
genomic library Lee and Kocher, 1996
were purchased from Research Genetics Ž
. Huntsville, AL, USA . These were screened for the ability to amplify alleles and
demonstrate polymorphism in the mapping family and in individuals from each of the five ‘‘pure’’ species used in the breeding program. PCRs were performed in 10 ml
Ž .
volumes in a PTC-100-96V Thermocycler MJR; Watertown, MA, USA . The reaction Ž
conditions were as follows: Gibco 1 = PCR buffer 20 mM Tris–HCl pH 8.0, 50 mM .
KCl , 2.4 mM MgCl , 160 mM each dNTP, 0.2 U Taq polymerase, 0.16 mM each
2
primer, and 10 ng of template DNA. An initial screen of the primers for amplification and polymorphism was done at 28 cycles of 948C for 1 min, 508C for 2 min, and 728C
for 1 min. The annealing temperature was adjusted for each locus as necessary to produce scorable amplification product.
2.2.3. AFLP protocol Ž
. The use of AFLP markers was based on a modification of Vos et al. 1995 . As
described below, visualization of bands in gels and data collection utilized a fluorescent scanning system.
Ž .
2.2.3.1. Digestion–ligation. Genomic DNA 200 ng was restricted for 1 h at 378C in a Ž
200-ml tube containing 2 U each of EcoRI and MseI New England Biolabs; Beverly, .
Ž MA, USA and 1 = NEBuffer 2 10 mM Tris–HCl, 10 mM MgCl , 50 mM NaCl, 1
2
. mM DTT, 100 mgrml BSA, pH 7.9 in a total volume of 16 ml. After digestion, 4 ml of
Ž a ligation mixture 20 pmol of MseI adapter, 2 pmol of EcoRI adapter, 1 mM ATP, 4 U
Ž .
. T4 DNA ligase New England Biolabs , 100 mgrml BSA, and 1 = NEBuffer 2 was
added. The reaction mixture was incubated at 258C for 12–16 h, and then diluted Ž
. 10-fold with 180 ml of TLE 10 mM Tris–HCl pH 8.0, 0.1 mM EDTA .
2.2.3.2. Pre-amplification. A pre-amplification step was used to reduce the number of possible bands to amplify, and to create a virtually unlimited supply of digestion–liga-
tion template. Primers complementary to each adapter sequence with a one-base extension were used to amplify 1r16 of the total restriction fragments created in the
digestion–ligation above. The template, 5 ml of diluted digestion–ligation product, was amplified using 30 ng each of Eco q G and Mse q C primers in a 20-ml reaction
Ž .
containing Gibco 1 = PCR buffer 20 mM Tris–HCl pH 8.0, 50 mM KCl , 1.5 mM MgCl , 200 mM of each dNTP, and 0.4 U of Taq polymerase. The template was
2
amplified for 24 cycles of: 948 for 30 s, 568C for 1 min, and 728C for 1 min. The PCR product was diluted 10-fold with 180 ml of TLE, and stored at 48C.
2.2.3.3. SelectiÕe amplification. It was determined empirically that a two-base extension on the fluorescein-labeled EcoRI primer and a three-base extension on the MseI primer
produced the optimum number of bands to be easily and unambiguously analyzed on a gel. Selective amplifications were carried out in a 20-ml reaction where 5 ml of diluted
pre-amplification product were amplified using the same reaction conditions as for the pre-amplification, except that the amount of primers used was 5 and 30 ng for the Eco
and Mse primer, respectively. Reactions were amplified with 30 s denaturing at 948C, followed by 30 s annealing at 658C and 1 min extension at 728C. This cycle was
repeated 11 more times reducing the annealing temperature 0.78C each cycle. The cycle was repeated a final 24 times with the annealing temperature held at a constant 568C.
2.2.4. Electrophoresis Ž
After amplification, an equal volume of loading buffer 98 formamide, 10 mM .
EDTA was added to the amplification products of both the microsatellite reactions and the AFLP reactions. The reactions were denatured at 958C for 2 min, then snap-cooled
Ž on ice for at least 5 min prior to loading on a 5 polyacrylamide 19:1 acrylamide:bi-
. Ž
. sacrylamide , 7.5 M urea denaturing gel 20 cm = 32 cm = 0.4 mm . Microsatellite
Ž .
products 48 samples; 2.5 ml each were loaded on the gel and run at 35 W for 25 min. Then another 48 reactions from the same locus were loaded and the gel was run for 30
Ž .
more minutes. AFLP products 48 samples were loaded and the gels were run at 35W Ž
. 508C between 1 and 1.5 h, depending on the fragment sizes for a particular primer
combination. 2.2.5. Scoring and analysis
Ž .
Each band on the AFLP gel was scored as band presence qrq,qry as one Ž
. allele, or absence yry as the alternate allele. Although it has been reported possible
Ž .
in other studies van Eck et al., 1995; Vos et al., 1995 , we did not feel confident Ž
. enough distinguishing differences in band intensities between homozygotes qrq and
Ž .
heterozygotes qry to score the AFLP bands as codominant markers. For all loci scored, the approach used was to score only loci for which one parent was scored as
band-present and the other parent was scored as band-absent. All loci scored were checked for segregation distortion using a x
2
for goodness of fit to a 1:1 genotype ratio Ž
. p - 0.05 . Twenty-five loci that showed a significant deviation were not included in
further analyses. Ž
Bands were sized using ‘‘FragmenNT Analysis’’ Molecular Dynamics; Sunnyvale, .
CA, USA and genotypes were scored by the presence or absence of bands by eye. Ž
. Linkage was analyzed using Mapmaker 3.0 Lander et al., 1987 . The LOD criterion for
determining linkage was taken as 3.0 and the map units used were Kosambi centimor- gans.
The linkage phase for some of the loci from parental fish used to make the cross was not known. Therefore, data were entered into Mapmaker first as the bands appeared on
Ž .
the gel band present s H, band absent s A , and as the reciprocal data set by changing A’s to H’s and H’s to A’s. This process created two identical but reciprocal sets of
linkage groups, so only one set was used for further study. Ž
Data also were organized using ‘‘Map Manager XP’’ Cudmore and Manly,
. http:rrmcbio.med.buffalo.edurmapmgr.html , which provides graphic display of the
genotypes of linked loci, and therefore enables visual inspection of recombination events. When double crossovers were observed in a small area, an indicator of possible
Ž .
genotyping errors Lincoln and Lander, 1992 , gels were reexamined and corrections were made in the data set if necessary. Linkage groups were drawn using the ‘‘Draw-
Ž .
Map’’ package Van Ooijen, 1994 .
3. Results
